Evaluating Lookup-Based Monocular Human Pose Tracking on the HumanEva Test Data
|
|
- Stanley Armstrong
- 3 years ago
- Views:
Transcription
1 Evaluating Lookup-Based Monocular Human Pose Tracking on the HumanEva Test Data Nicholas R. Howe Computer Science Department Smith College Northampton, MA Abstract This work presents an evaluation of several lookup-based methods for recovering three-dimensional human pose from monocular video sequences. The methods themselves are largely described elsewhere [1, 2], although the work presented here incorporates a few minor enhancements. The primary contribution of this work is the evaluation of the results on a data set with ground truth available, which allows for quantitative comparisons with other techniques. Methods relying upon silhouettes produced via background subtraction tend to act as a straw man in relation to the current state of the art; many recently proposed techniques work without reliance upon background subtraction and cite this feature as one of their advantages. Without disputing such reasonable claims, this work seeks to push the envelope for background-subtraction methods as far as possible. The goal of this effort is to provide a challenging baseline against which the performance of various alternatives may be assessed. 1 Background Subtraction Given the commitment to background subtraction made here, the quality of the extracted silhouettes will strongly affect the result. Although the reconstruction methods may cope to some extent with noisy silhouettes, for the strongest comparison the silhouettes should be nearly error-free. Fortunately, recent work has demonstrated that graph-based techniques can extract high-quality silhouettes both reliably and quickly in most cases [5]. The foreground segmentation adopted here is based upon a different implementation with a similar philosophy, as detailed in the items below. The trim mean gives a robust Gaussian model of the background color at each pixel. For clips where the subject moves around sufficiently, the background model can be estimated directly from the action video. For a given frame, the number of standard deviations from the background color model at each pixel guides the foreground segmentation. For color images, separate models are developed for hue, saturation, and value at each pixel. A linear combination of the results of the three models weights each one according to its reliability. (The hue channel shows greater noise even after normalizing by the standard deviation, and is consequently weighted less than the other two.) To mitigate shadows, the model forgives luminosity decreases of up to τ s from the computed background luminosity. This accounts for possible darkening due to shadows, and typically improves the segmentation where the subject s feet meet the floor. Occasionally, it may improperly label foreground regions as background if their color is slightly darker than the background region they occlude.
2 Figure 1: Example of typical foreground segmentation result. The precise boundaries and separation of body parts make further pose recovery steps easier. Nevertheless, errors can occur where there is poor contrast between the subject and the background, as in the darker portion of the left shoe. The minimum cut on a graph constructed from the image gives the foreground segmentation. Both four- and eight-connected neighbor edges are included in the graph, with weaker links to diagonal neighbors so that the solution favors neither straight nor diagonal boundaries. The graph omits neighbor edges where gradients appear in the frame that are not present in the background image. This strategy biases the foreground segmentation to follow object boundaries in the image. Figure 1 shows an example of a segmented frame. 2 Method The method employed in this work is called lookup-based because it recovers 3D pose by retrieving perceptually similar examples from a previously computed collection. A small set of poses selected for each frame forms the basis of the next processing stage, which imposes a temporal Markov condition to select a sequence of poses with small momentum changes and high fidelity to optical flow observations. (This stitching process begins with 10 to 50 candidates per frame, depending on their similarity score.) Finally, smoothing the pose trajectories over time eliminates any remaining jerkiness. The lookup stage may use any number of retrieval methods to generate the initial candidate pool, and this work combines several. It begins with three basic similarity measures, computed from the image silhouette and optical flow: moments of the image optical flow[2], turning angle[1], and half-chamfer distance. (Unlike full chamfer, half-chamfer can be computed in a small number of calculations.) The candidate pool includes poses that score the top rankings on all three measures, on the flow moments alone, and on the two shape measures together. Some are constrained to lie near poses in the candidate pool from the previous frame, ensuring continuity; a handful are chosen without this restraint to allow for recovery from errors. Although up to 65 candidates may be selected, overlap between the different retrieval modes means that the actual number will usually be lower. These candidates are registered to the image frame using a full chamfer match on the silhouette, and the most successful registrations compete for selection in the next stage. The temporal chaining step employs a two-state stochastic chain model (sometimes called a secondorder hidden Markov model) to find the sequence of states minimizing a global energy function. The terms in this global function represent three balanced concerns: the match of each candidate pose to its frame, the momentum change embodied in each three-frame section (including estimates of both linear and angular momentum), and the match of the implied optical flow to the flow observed in the image. Previous work on shorter video clips has suggested that including the latter two constraints in the energy function can prevent limb-swapping errors that sometimes appear when one leg crosses in front of another[2]. The results reported here suggest that they are not entirely successful, and further study of this issue may be warranted. If limb-swapping is suppressed via other means (for example, a state transition model for walking [3]), then a simpler formulation is possible that omits flow comparisons and employs only a first-order HMM. Nevertheless, the results presented here all use the more complex model described initially.
3 The smoothing stage of processing is mostly cosmetic, and can in some cases increase the error in comparison to ground truth. The smoothing treats individual pose parameters as one-dimensional functions over time, and removes any high-frequency variation. Previous work employed a method that modeled overlapping subsections via quadratics and blended the results together[1]. The results here instead perform a low-pass smoothing by zeroing high-frequency components of the DCT, but the result is quite similar. 3 Experiments The HumanEva test set [4] forms the basis of these experiments. This abstract presents results only for a few selections from the walking sequences. We expect to process additional sequences as time permits. Processing proceeds in several stages. The first stage computes a foreground segmentation and optical flow field for each frame. The second stage uses the foreground silhouette and the moments of the optical flow to retrieve matching silhouettes with known poses from a collection derived from the CMU motion capture library. For greater reliability, this stage performs several retrievals employing different combinations of similarity measures: flow moments and turning angle and halfchamfer matching all together, flow moments alone, and turning angle and half-chamfer matching together. In addition, some of the retrieved poses are constrained to remain close to the best matches from the previous frame, to ensure continuity in difficult regions. All the retrieved poses (at most 65) are pooled together, with any duplicates removed, and registered to the image frame. The third stage applies a second-order Markov chain to identify a temporal sequence of poses with low-energy transitions. Good sequences have small positional changes from frame to frame, and their calculated flow matches the flow observed in the frame. A final stage can smooth the temporal pose sequence to suppress jerkiness. However, this step can actually increase the error as compared to ground truth, so it is omitted here. Reconstructed poses are registered to the video frame and therefore comparable to the known ground truth in the HumanEva suite. The analysis below computes the mean error over all the evaluation points, plus the mean error for each point over all the frames. 1 The results reveal one glaring shortcoming of the method. Silhouette-based reconstructions are known to suffer from an ambiguity where the pose is inverted along the line-of-sight axis and the left and right sides of the body are exchanged. Such a transformation produces a physically reasonable pose with exactly the same silhouette as the original. In previous work the inclusion of flow data suppressed this problem [2], but it occurs often in the HumanEva sequences despite the flow constraints. One easy way to solve this problem would be to introduce a walking-specific motion model that requires alternating strides, as some have done[3], but this would compromise the generality of the lookup-based approach. Although this matter clearly invites further research, for the moment the results below present both the raw error (which may be quite high due to inversion of the recovered pose) and the minimum error chosen between the recovered pose and its inversion. Labels in the text will endeavor to indicate clearly which is which. Indicative results appear here for the S3/Walking 1 (C2) and S3/Walking 1 (BW2) videos. The choice of subject and camera are not particularly significant, except that the background subtraction contains no extreme errors. Figures 3 and 2 show some statistics on the results, organized by frame and by evaluation point. The overall mean errors after allowing for pose inversion are 14 and 11 pixels, respectively. However, the figures reveal considerable variation. The individual frame errors show a standard deviation of nearly 9 pixels. The mean joint errors ranged from a low of 8 pixels for the pelvis to as much as 29 pixels for the extremities of the arms. The legs, despite their greater length, show lower mean error than the arms because they are more often visible in the silhouette. The results presented in these figures have not been optimized in any way from the silhouettes selected by the optimal Markov chain. Naturally, the retrieved poses will not fit the observed motion exactly, and should allow further improvement via numeric optimization. A very preliminary experiment indicates the truth of this hypothesis: one pass of frame-by-frame optimization 1 For the videos, visual inspection revealed what appeared to be an error in the ground truth data for the top of the head over many frames; consequently this point was omitted from the frame-by-frame analysis.
4 Figure 2: Summary of results for the S3/Walking 1 (BW2) sequence. At left, the red dotted line shows the error of the reconstructed pose. The solid line shows the minimum error of the reconstructed pose or its inversion (cyan dotted). Higher frame error in this sequence appears correlated with sideways walking. At right, the bars show the mean error for each evaluation point over all the frames. They show that error increases towards the extremities. The large error in the first point incorporates apparently erroneous data. Figure 3: Summary of results for the S3/Walking 1 (C2) sequence. At left, the red dotted line shows the error of the reconstructed pose. The solid line shows the minimum error of the reconstructed pose or its inversion (cyan dotted). The larger error around frames may be correlated with instability in the background subtraction. At right, the bars show the mean error for each evaluation point over all the frames. They show that error increases towards the extremities. The large error in the first point incorporates the apparently erroneous data. using chamfer-distance and temporal smoothness criteria reduces the mean ground truth error in the S3/Walking 1 (BW2) video from 11 to 10 pixels. 4 Conclusion The numeric results show considerable room for improvement, either through refinements of the silhouette lookup method or through completely different approaches. These results are presented in the hopes that they will add to understanding of the state of the art, and spur further research in the field. It is our intention to add results on additional HumanEva sequences as soon as possible. Acknowledgment This manuscript is based upon work supported by the National Science Foundation under Grant No. IIS The data used in this project was obtained from mocap.cs.cmu.edu. The database was created with funding from NSF EIA
5 References [1] N. Howe. Silhouette lookup for automatic pose tracking. In IEEE Workshop on Articulated and Nonrigid Motion, [2] N. Howe. Flow lookup and biological motion perception. In International Conference on Image Processing, [3] X. Lan and D. Huttenlocher. A unified spatio-temporal articulated model for tracking. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition, volume I, pages , [4] L. Sigal and M. Black. HumanEva: Synchronized video and motion capture dataset for evaluation of articulated human motion. Technical Report CS-06-08, Brown University, September [5] Y. Sun, B. Yuan, Z. Miao, and C. Wan. Better foreground segmentation for static cameras via new energy form and dynamic graph-cut. In ICPR (4), pages 49 52, 2006.
Manual for Human Tracking Software
Manual for Human Tracking Software Alexandru Balan Department of Computer Science Brown University Providence, Rhode Island 02912 alb@cs.brown.edu Version 1.0 November 26, 2005 1. Abstract A novel human
More informationModel-Based 3D Human Motion Capture Using Global-Local Particle Swarm Optimizations
Model-Based 3D Human Motion Capture Using Global-Local Particle Swarm Optimizations Tomasz Krzeszowski, Bogdan Kwolek, and Konrad Wojciechowski Abstract. We present an approach for tracking the articulated
More informationVEHICLE LOCALISATION AND CLASSIFICATION IN URBAN CCTV STREAMS
VEHICLE LOCALISATION AND CLASSIFICATION IN URBAN CCTV STREAMS Norbert Buch 1, Mark Cracknell 2, James Orwell 1 and Sergio A. Velastin 1 1. Kingston University, Penrhyn Road, Kingston upon Thames, KT1 2EE,
More informationBlind Deconvolution of Barcodes via Dictionary Analysis and Wiener Filter of Barcode Subsections
Blind Deconvolution of Barcodes via Dictionary Analysis and Wiener Filter of Barcode Subsections Maximilian Hung, Bohyun B. Kim, Xiling Zhang August 17, 2013 Abstract While current systems already provide
More informationSpeed Performance Improvement of Vehicle Blob Tracking System
Speed Performance Improvement of Vehicle Blob Tracking System Sung Chun Lee and Ram Nevatia University of Southern California, Los Angeles, CA 90089, USA sungchun@usc.edu, nevatia@usc.edu Abstract. A speed
More informationVision based Vehicle Tracking using a high angle camera
Vision based Vehicle Tracking using a high angle camera Raúl Ignacio Ramos García Dule Shu gramos@clemson.edu dshu@clemson.edu Abstract A vehicle tracking and grouping algorithm is presented in this work
More informationCanny Edge Detection
Canny Edge Detection 09gr820 March 23, 2009 1 Introduction The purpose of edge detection in general is to significantly reduce the amount of data in an image, while preserving the structural properties
More informationSupporting Online Material for
www.sciencemag.org/cgi/content/full/313/5786/504/dc1 Supporting Online Material for Reducing the Dimensionality of Data with Neural Networks G. E. Hinton* and R. R. Salakhutdinov *To whom correspondence
More informationColorado School of Mines Computer Vision Professor William Hoff
Professor William Hoff Dept of Electrical Engineering &Computer Science http://inside.mines.edu/~whoff/ 1 Introduction to 2 What is? A process that produces from images of the external world a description
More informationThe Visual Internet of Things System Based on Depth Camera
The Visual Internet of Things System Based on Depth Camera Xucong Zhang 1, Xiaoyun Wang and Yingmin Jia Abstract The Visual Internet of Things is an important part of information technology. It is proposed
More informationBCC Multi Stripe Wipe
BCC Multi Stripe Wipe The BCC Multi Stripe Wipe is a similar to a Horizontal or Vertical Blind wipe. It offers extensive controls to randomize the stripes parameters. The following example shows a Multi
More informationA Quantitative Evaluation of Video-based 3D Person Tracking
A Quantitative Evaluation of Video-based 3D Person Tracking Alexandru O. Bălan Leonid Sigal Michael J. Black Department of Computer Science, Brown University, Providence, RI 2912, USA {alb, ls, black}@cs.brown.edu
More informationTracking and Recognition in Sports Videos
Tracking and Recognition in Sports Videos Mustafa Teke a, Masoud Sattari b a Graduate School of Informatics, Middle East Technical University, Ankara, Turkey mustafa.teke@gmail.com b Department of Computer
More informationFeature Tracking and Optical Flow
02/09/12 Feature Tracking and Optical Flow Computer Vision CS 543 / ECE 549 University of Illinois Derek Hoiem Many slides adapted from Lana Lazebnik, Silvio Saverse, who in turn adapted slides from Steve
More informationObject tracking & Motion detection in video sequences
Introduction Object tracking & Motion detection in video sequences Recomended link: http://cmp.felk.cvut.cz/~hlavac/teachpresen/17compvision3d/41imagemotion.pdf 1 2 DYNAMIC SCENE ANALYSIS The input to
More informationPractical Tour of Visual tracking. David Fleet and Allan Jepson January, 2006
Practical Tour of Visual tracking David Fleet and Allan Jepson January, 2006 Designing a Visual Tracker: What is the state? pose and motion (position, velocity, acceleration, ) shape (size, deformation,
More informationStatic Environment Recognition Using Omni-camera from a Moving Vehicle
Static Environment Recognition Using Omni-camera from a Moving Vehicle Teruko Yata, Chuck Thorpe Frank Dellaert The Robotics Institute Carnegie Mellon University Pittsburgh, PA 15213 USA College of Computing
More informationColour Image Segmentation Technique for Screen Printing
60 R.U. Hewage and D.U.J. Sonnadara Department of Physics, University of Colombo, Sri Lanka ABSTRACT Screen-printing is an industry with a large number of applications ranging from printing mobile phone
More informationReal-Time Tracking of Pedestrians and Vehicles
Real-Time Tracking of Pedestrians and Vehicles N.T. Siebel and S.J. Maybank. Computational Vision Group Department of Computer Science The University of Reading Reading RG6 6AY, England Abstract We present
More informationEdge tracking for motion segmentation and depth ordering
Edge tracking for motion segmentation and depth ordering P. Smith, T. Drummond and R. Cipolla Department of Engineering University of Cambridge Cambridge CB2 1PZ,UK {pas1001 twd20 cipolla}@eng.cam.ac.uk
More informationBinocular Vision and The Perception of Depth
Binocular Vision and The Perception of Depth Visual Perception How one visually interprets a scene 4 forms of perception to be studied: Depth Color Temporal Motion Depth Perception How does one determine
More informationFreehand Sketching. Sections
3 Freehand Sketching Sections 3.1 Why Freehand Sketches? 3.2 Freehand Sketching Fundamentals 3.3 Basic Freehand Sketching 3.4 Advanced Freehand Sketching Key Terms Objectives Explain why freehand sketching
More informationReconstructing 3D Pose and Motion from a Single Camera View
Reconstructing 3D Pose and Motion from a Single Camera View R Bowden, T A Mitchell and M Sarhadi Brunel University, Uxbridge Middlesex UB8 3PH richard.bowden@brunel.ac.uk Abstract This paper presents a
More informationMASKS & CHANNELS WORKING WITH MASKS AND CHANNELS
MASKS & CHANNELS WORKING WITH MASKS AND CHANNELS Masks let you isolate and protect parts of an image. When you create a mask from a selection, the area not selected is masked or protected from editing.
More informationAn Active Head Tracking System for Distance Education and Videoconferencing Applications
An Active Head Tracking System for Distance Education and Videoconferencing Applications Sami Huttunen and Janne Heikkilä Machine Vision Group Infotech Oulu and Department of Electrical and Information
More informationOptical Flow. Shenlong Wang CSC2541 Course Presentation Feb 2, 2016
Optical Flow Shenlong Wang CSC2541 Course Presentation Feb 2, 2016 Outline Introduction Variation Models Feature Matching Methods End-to-end Learning based Methods Discussion Optical Flow Goal: Pixel motion
More informationImage Segmentation and Registration
Image Segmentation and Registration Dr. Christine Tanner (tanner@vision.ee.ethz.ch) Computer Vision Laboratory, ETH Zürich Dr. Verena Kaynig, Machine Learning Laboratory, ETH Zürich Outline Segmentation
More informationMean-Shift Tracking with Random Sampling
1 Mean-Shift Tracking with Random Sampling Alex Po Leung, Shaogang Gong Department of Computer Science Queen Mary, University of London, London, E1 4NS Abstract In this work, boosting the efficiency of
More informationMachine vision systems - 2
Machine vision systems Problem definition Image acquisition Image segmentation Connected component analysis Machine vision systems - 1 Problem definition Design a vision system to see a flat world Page
More informationCHAPTER 6 TEXTURE ANIMATION
CHAPTER 6 TEXTURE ANIMATION 6.1. INTRODUCTION Animation is the creating of a timed sequence or series of graphic images or frames together to give the appearance of continuous movement. A collection of
More informationPart-Based Recognition
Part-Based Recognition Benedict Brown CS597D, Fall 2003 Princeton University CS 597D, Part-Based Recognition p. 1/32 Introduction Many objects are made up of parts It s presumably easier to identify simple
More informationJPEG compression of monochrome 2D-barcode images using DCT coefficient distributions
Edith Cowan University Research Online ECU Publications Pre. JPEG compression of monochrome D-barcode images using DCT coefficient distributions Keng Teong Tan Hong Kong Baptist University Douglas Chai
More informationWhitepaper. Image stabilization improving camera usability
Whitepaper Image stabilization improving camera usability Table of contents 1. Introduction 3 2. Vibration Impact on Video Output 3 3. Image Stabilization Techniques 3 3.1 Optical Image Stabilization 3
More informationTracking Groups of Pedestrians in Video Sequences
Tracking Groups of Pedestrians in Video Sequences Jorge S. Marques Pedro M. Jorge Arnaldo J. Abrantes J. M. Lemos IST / ISR ISEL / IST ISEL INESC-ID / IST Lisbon, Portugal Lisbon, Portugal Lisbon, Portugal
More informationSubspace Analysis and Optimization for AAM Based Face Alignment
Subspace Analysis and Optimization for AAM Based Face Alignment Ming Zhao Chun Chen College of Computer Science Zhejiang University Hangzhou, 310027, P.R.China zhaoming1999@zju.edu.cn Stan Z. Li Microsoft
More informationLimits and Possibilities of Markerless Human Motion Estimation
Limits and Possibilities of Markerless Human Motion Estimation Bodo Rosenhahn Universität Hannover Motion Capture Wikipedia (MoCap): Approaches for recording and analyzing Human motions Markerless Motion
More informationDrawing Lines with Pixels. Joshua Scott March 2012
Drawing Lines with Pixels Joshua Scott March 2012 1 Summary Computers draw lines and circles during many common tasks, such as using an image editor. But how does a computer know which pixels to darken
More informationCS231M Project Report - Automated Real-Time Face Tracking and Blending
CS231M Project Report - Automated Real-Time Face Tracking and Blending Steven Lee, slee2010@stanford.edu June 6, 2015 1 Introduction Summary statement: The goal of this project is to create an Android
More informationTracking And Object Classification For Automated Surveillance
Tracking And Object Classification For Automated Surveillance Omar Javed and Mubarak Shah Computer Vision ab, University of Central Florida, 4000 Central Florida Blvd, Orlando, Florida 32816, USA {ojaved,shah}@cs.ucf.edu
More informationBernice E. Rogowitz and Holly E. Rushmeier IBM TJ Watson Research Center, P.O. Box 704, Yorktown Heights, NY USA
Are Image Quality Metrics Adequate to Evaluate the Quality of Geometric Objects? Bernice E. Rogowitz and Holly E. Rushmeier IBM TJ Watson Research Center, P.O. Box 704, Yorktown Heights, NY USA ABSTRACT
More informationImplementation of Canny Edge Detector of color images on CELL/B.E. Architecture.
Implementation of Canny Edge Detector of color images on CELL/B.E. Architecture. Chirag Gupta,Sumod Mohan K cgupta@clemson.edu, sumodm@clemson.edu Abstract In this project we propose a method to improve
More informationSolidWorks Implementation Guides. Sketching Concepts
SolidWorks Implementation Guides Sketching Concepts Sketching in SolidWorks is the basis for creating features. Features are the basis for creating parts, which can be put together into assemblies. Sketch
More informationQuantitative vs. Categorical Data: A Difference Worth Knowing Stephen Few April 2005
Quantitative vs. Categorical Data: A Difference Worth Knowing Stephen Few April 2005 When you create a graph, you step through a series of choices, including which type of graph you should use and several
More informationOverview. Raster Graphics and Color. Overview. Display Hardware. Liquid Crystal Display (LCD) Cathode Ray Tube (CRT)
Raster Graphics and Color Greg Humphreys CS445: Intro Graphics University of Virginia, Fall 2004 Color models Color models Display Hardware Video display devices Cathode Ray Tube (CRT) Liquid Crystal Display
More informationDESCRIPTIVE STATISTICS. The purpose of statistics is to condense raw data to make it easier to answer specific questions; test hypotheses.
DESCRIPTIVE STATISTICS The purpose of statistics is to condense raw data to make it easier to answer specific questions; test hypotheses. DESCRIPTIVE VS. INFERENTIAL STATISTICS Descriptive To organize,
More informationLinear Programming. Solving LP Models Using MS Excel, 18
SUPPLEMENT TO CHAPTER SIX Linear Programming SUPPLEMENT OUTLINE Introduction, 2 Linear Programming Models, 2 Model Formulation, 4 Graphical Linear Programming, 5 Outline of Graphical Procedure, 5 Plotting
More informationMulti-Zone Adjustment
Written by Jonathan Sachs Copyright 2008 Digital Light & Color Introduction Picture Window s 2-Zone Adjustment and3-zone Adjustment transformations are powerful image enhancement tools designed for images
More informationROBUST VEHICLE TRACKING IN VIDEO IMAGES BEING TAKEN FROM A HELICOPTER
ROBUST VEHICLE TRACKING IN VIDEO IMAGES BEING TAKEN FROM A HELICOPTER Fatemeh Karimi Nejadasl, Ben G.H. Gorte, and Serge P. Hoogendoorn Institute of Earth Observation and Space System, Delft University
More informationMulti-view Intelligent Vehicle Surveillance System
Multi-view Intelligent Vehicle Surveillance System S. Denman, C. Fookes, J. Cook, C. Davoren, A. Mamic, G. Farquharson, D. Chen, B. Chen and S. Sridharan Image and Video Research Laboratory Queensland
More informationAutoCAD 2009. New Icon Quick Reference
AutoCAD 2009 New Quick Reference Contents Chapter 1 New Quick Reference..................... 1 Toolbars................................... 1 3D Navigation Toolbar........................ 1 CAD Standards
More informationAutomatic Labeling of Lane Markings for Autonomous Vehicles
Automatic Labeling of Lane Markings for Autonomous Vehicles Jeffrey Kiske Stanford University 450 Serra Mall, Stanford, CA 94305 jkiske@stanford.edu 1. Introduction As autonomous vehicles become more popular,
More informationData representation and analysis in Excel
Page 1 Data representation and analysis in Excel Let s Get Started! This course will teach you how to analyze data and make charts in Excel so that the data may be represented in a visual way that reflects
More informationMeasuring Line Edge Roughness: Fluctuations in Uncertainty
Tutor6.doc: Version 5/6/08 T h e L i t h o g r a p h y E x p e r t (August 008) Measuring Line Edge Roughness: Fluctuations in Uncertainty Line edge roughness () is the deviation of a feature edge (as
More informationOBJECT TRACKING USING LOG-POLAR TRANSFORMATION
OBJECT TRACKING USING LOG-POLAR TRANSFORMATION A Thesis Submitted to the Gradual Faculty of the Louisiana State University and Agricultural and Mechanical College in partial fulfillment of the requirements
More informationHow To Analyze Ball Blur On A Ball Image
Single Image 3D Reconstruction of Ball Motion and Spin From Motion Blur An Experiment in Motion from Blur Giacomo Boracchi, Vincenzo Caglioti, Alessandro Giusti Objective From a single image, reconstruct:
More informationTracking performance evaluation on PETS 2015 Challenge datasets
Tracking performance evaluation on PETS 2015 Challenge datasets Tahir Nawaz, Jonathan Boyle, Longzhen Li and James Ferryman Computational Vision Group, School of Systems Engineering University of Reading,
More informationA Learning Based Method for Super-Resolution of Low Resolution Images
A Learning Based Method for Super-Resolution of Low Resolution Images Emre Ugur June 1, 2004 emre.ugur@ceng.metu.edu.tr Abstract The main objective of this project is the study of a learning based method
More informationWhat is Visualization? Information Visualization An Overview. Information Visualization. Definitions
What is Visualization? Information Visualization An Overview Jonathan I. Maletic, Ph.D. Computer Science Kent State University Visualize/Visualization: To form a mental image or vision of [some
More informationScanners and How to Use Them
Written by Jonathan Sachs Copyright 1996-1999 Digital Light & Color Introduction A scanner is a device that converts images to a digital file you can use with your computer. There are many different types
More informationARC 3D Webservice How to transform your images into 3D models. Maarten Vergauwen info@arc3d.be
ARC 3D Webservice How to transform your images into 3D models Maarten Vergauwen info@arc3d.be Overview What is it? How does it work? How do you use it? How to record images? Problems, tips and tricks Overview
More informationA Prototype For Eye-Gaze Corrected
A Prototype For Eye-Gaze Corrected Video Chat on Graphics Hardware Maarten Dumont, Steven Maesen, Sammy Rogmans and Philippe Bekaert Introduction Traditional webcam video chat: No eye contact. No extensive
More informationChapter 5: Working with contours
Introduction Contoured topographic maps contain a vast amount of information about the three-dimensional geometry of the land surface and the purpose of this chapter is to consider some of the ways in
More informationVideo compression: Performance of available codec software
Video compression: Performance of available codec software Introduction. Digital Video A digital video is a collection of images presented sequentially to produce the effect of continuous motion. It takes
More informationSession 7 Bivariate Data and Analysis
Session 7 Bivariate Data and Analysis Key Terms for This Session Previously Introduced mean standard deviation New in This Session association bivariate analysis contingency table co-variation least squares
More informationAre Those Sunspots Really on the Sun?
Are Those Sunspots Really on the Sun? Summary of Activity: Students will acquire solar images (or draw sunspots), and record coordinates of sunspots. They will calculate and plot their apparent movement
More informationUsing Image J to Measure the Brightness of Stars (Written by Do H. Kim)
Using Image J to Measure the Brightness of Stars (Written by Do H. Kim) What is Image J? Image J is Java-based image processing program developed at the National Institutes of Health. Image J runs on everywhere,
More informationACE: After Effects CC
Adobe Training Services Exam Guide ACE: After Effects CC Adobe Training Services provides this exam guide to help prepare partners, customers, and consultants who are actively seeking accreditation as
More informationChapter 1. Creating Sketches in. the Sketch Mode-I. Evaluation chapter. Logon to www.cadcim.com for more details. Learning Objectives
Chapter 1 Creating Sketches in Learning Objectives the Sketch Mode-I After completing this chapter you will be able to: Use various tools to create a geometry. Dimension a sketch. Apply constraints to
More informationINSTRUCTOR WORKBOOK Quanser Robotics Package for Education for MATLAB /Simulink Users
INSTRUCTOR WORKBOOK for MATLAB /Simulink Users Developed by: Amir Haddadi, Ph.D., Quanser Peter Martin, M.A.SC., Quanser Quanser educational solutions are powered by: CAPTIVATE. MOTIVATE. GRADUATE. PREFACE
More informationReflection and Refraction
Equipment Reflection and Refraction Acrylic block set, plane-concave-convex universal mirror, cork board, cork board stand, pins, flashlight, protractor, ruler, mirror worksheet, rectangular block worksheet,
More informationHead-Coupled Perspective
Head-Coupled Perspective Introduction Head-Coupled Perspective (HCP) refers to a technique of rendering a scene that takes into account the position of the viewer relative to the display. As a viewer moves
More informationVIRGINIA WESTERN COMMUNITY COLLEGE
36T Revised Fall 2015 Cover Page 36TITD 112 21TDesigning Web Page Graphics Program Head: Debbie Yancey Revised: Fall 2015 Dean s Review: Deborah Yancey Dean 21T Lab/Recitation Revised Fall 2015 None ITD
More informationEpipolar Geometry. Readings: See Sections 10.1 and 15.6 of Forsyth and Ponce. Right Image. Left Image. e(p ) Epipolar Lines. e(q ) q R.
Epipolar Geometry We consider two perspective images of a scene as taken from a stereo pair of cameras (or equivalently, assume the scene is rigid and imaged with a single camera from two different locations).
More informationHow To Fix Out Of Focus And Blur Images With A Dynamic Template Matching Algorithm
IJSTE - International Journal of Science Technology & Engineering Volume 1 Issue 10 April 2015 ISSN (online): 2349-784X Image Estimation Algorithm for Out of Focus and Blur Images to Retrieve the Barcode
More informationEFX Keying/Alpha plugins for After Effects
EFX Keying/Alpha plugins for After Effects Here you'll find description of the plugins developed to help select pixels based on many criteria. Also after selection, there are often things you want to do
More informationAdvantages of Auto-tuning for Servo-motors
Advantages of for Servo-motors Executive summary The same way that 2 years ago computer science introduced plug and play, where devices would selfadjust to existing system hardware, industrial motion control
More informationCS 4204 Computer Graphics
CS 4204 Computer Graphics Computer Animation Adapted from notes by Yong Cao Virginia Tech 1 Outline Principles of Animation Keyframe Animation Additional challenges in animation 2 Classic animation Luxo
More informationCollaborative Data Analysis on Wall Displays
Collaborative Data Analysis on Wall Displays Challenges for Visualization Petra Isenberg (petra.isenberg@inria.fr) Anastasia Bezerianos (anastasia.bezerianos@lri.fr) 2 [source: The Diverse and Exploding
More informationUrban Vehicle Tracking using a Combined 3D Model Detector and Classifier
Urban Vehicle Tracing using a Combined 3D Model Detector and Classifier Norbert Buch, Fei Yin, James Orwell, Dimitrios Maris and Sergio A. Velastin Digital Imaging Research Centre, Kingston University,
More informationBLIND SOURCE SEPARATION OF SPEECH AND BACKGROUND MUSIC FOR IMPROVED SPEECH RECOGNITION
BLIND SOURCE SEPARATION OF SPEECH AND BACKGROUND MUSIC FOR IMPROVED SPEECH RECOGNITION P. Vanroose Katholieke Universiteit Leuven, div. ESAT/PSI Kasteelpark Arenberg 10, B 3001 Heverlee, Belgium Peter.Vanroose@esat.kuleuven.ac.be
More informationLimitations of Human Vision. What is computer vision? What is computer vision (cont d)?
What is computer vision? Limitations of Human Vision Slide 1 Computer vision (image understanding) is a discipline that studies how to reconstruct, interpret and understand a 3D scene from its 2D images
More informationProcessing the Image or Can you Believe what you see? Light and Color for Nonscientists PHYS 1230
Processing the Image or Can you Believe what you see? Light and Color for Nonscientists PHYS 1230 Optical Illusions http://www.michaelbach.de/ot/mot_mib/index.html Vision We construct images unconsciously
More informationFalse alarm in outdoor environments
Accepted 1.0 Savantic letter 1(6) False alarm in outdoor environments Accepted 1.0 Savantic letter 2(6) Table of contents Revision history 3 References 3 1 Introduction 4 2 Pre-processing 4 3 Detection,
More informationHighlight Recovery in Camera Raw
WHITEPAPER A little-known capability of the Adobe Camera Raw plug-in is highlight recovery in the nearly overexposed highlights of a raw digital capture. Normally, when one or two of the channels in an
More informationCS 534: Computer Vision 3D Model-based recognition
CS 534: Computer Vision 3D Model-based recognition Ahmed Elgammal Dept of Computer Science CS 534 3D Model-based Vision - 1 High Level Vision Object Recognition: What it means? Two main recognition tasks:!
More informationMEASURES OF VARIATION
NORMAL DISTRIBTIONS MEASURES OF VARIATION In statistics, it is important to measure the spread of data. A simple way to measure spread is to find the range. But statisticians want to know if the data are
More information6 Space Perception and Binocular Vision
Space Perception and Binocular Vision Space Perception and Binocular Vision space perception monocular cues to 3D space binocular vision and stereopsis combining depth cues monocular/pictorial cues cues
More informationWii Remote Calibration Using the Sensor Bar
Wii Remote Calibration Using the Sensor Bar Alparslan Yildiz Abdullah Akay Yusuf Sinan Akgul GIT Vision Lab - http://vision.gyte.edu.tr Gebze Institute of Technology Kocaeli, Turkey {yildiz, akay, akgul}@bilmuh.gyte.edu.tr
More informationCharacter Image Patterns as Big Data
22 International Conference on Frontiers in Handwriting Recognition Character Image Patterns as Big Data Seiichi Uchida, Ryosuke Ishida, Akira Yoshida, Wenjie Cai, Yaokai Feng Kyushu University, Fukuoka,
More informationCharacter Animation from 2D Pictures and 3D Motion Data ALEXANDER HORNUNG, ELLEN DEKKERS, and LEIF KOBBELT RWTH-Aachen University
Character Animation from 2D Pictures and 3D Motion Data ALEXANDER HORNUNG, ELLEN DEKKERS, and LEIF KOBBELT RWTH-Aachen University Presented by: Harish CS-525 First presentation Abstract This article presents
More informationExperiment 3: Magnetic Fields of a Bar Magnet and Helmholtz Coil
MASSACHUSETTS INSTITUTE OF TECHNOLOGY Department of Physics 8.02 Spring 2009 Experiment 3: Magnetic Fields of a Bar Magnet and Helmholtz Coil OBJECTIVES 1. To learn how to visualize magnetic field lines
More informationInteractive Offline Tracking for Color Objects
Interactive Offline Tracking for Color Objects Yichen Wei Jian Sun Xiaoou Tang Heung-Yeung Shum Microsoft Research Asia, Beijing, China {yichenw,jiansun,xitang,hshum}@microsoft.com Abstract In this paper,
More informationGraphic Design. Background: The part of an artwork that appears to be farthest from the viewer, or in the distance of the scene.
Graphic Design Active Layer- When you create multi layers for your images the active layer, or the only one that will be affected by your actions, is the one with a blue background in your layers palette.
More informationIntroduction. Appendix D Mathematical Induction D1
Appendix D Mathematical Induction D D Mathematical Induction Use mathematical induction to prove a formula. Find a sum of powers of integers. Find a formula for a finite sum. Use finite differences to
More informationClassifying Manipulation Primitives from Visual Data
Classifying Manipulation Primitives from Visual Data Sandy Huang and Dylan Hadfield-Menell Abstract One approach to learning from demonstrations in robotics is to make use of a classifier to predict if
More informationExperiment 3: Magnetic Fields of a Bar Magnet and Helmholtz Coil
MASSACHUSETTS INSTITUTE OF TECHNOLOGY Department of Physics 8.02 Spring 2006 Experiment 3: Magnetic Fields of a Bar Magnet and Helmholtz Coil OBJECTIVES 1. To learn how to visualize magnetic field lines
More informationBeyond Built-in: Why a Better Webcam Matters
Whitepaper: Beyond Built-in: Why a Better Webcam Matters How to Uplevel Your Ability to Connect, Communicate and Collaborate Using Your Laptop or PC Introduction The ability to virtually communicate and
More informationAn Iterative Image Registration Technique with an Application to Stereo Vision
An Iterative Image Registration Technique with an Application to Stereo Vision Bruce D. Lucas Takeo Kanade Computer Science Department Carnegie-Mellon University Pittsburgh, Pennsylvania 15213 Abstract
More informationSTRESS FRACTURE ON A BOEING 757 : OBSERVATIONS BY A GEOLOGIST May 21, 2012
STRESS FRACTURE ON A BOEING 757 : OBSERVATIONS BY A GEOLOGIST Richard R. Donofrio May 21, 2012 Oklahoma City, OK As geologists we are accustomed to observing rock fractures in outcrops, cores, and elsewhere,
More information